CN116319549B

CN116319549B - Distributed flow scheduling method and device

Info

Publication number: CN116319549B
Application number: CN202310572187.1A
Authority: CN
Inventors: 蒋文栋
Original assignee: New H3C Technologies Co Ltd
Current assignee: New H3C Technologies Co Ltd
Priority date: 2023-05-18
Filing date: 2023-05-18
Publication date: 2023-08-18
Anticipated expiration: 2043-05-18
Also published as: CN116319549A

Abstract

The embodiment of the invention provides a distributed flow scheduling method and a device, which are applied to head node equipment in a SRv6 network, wherein the method comprises the following steps: when detecting that the current SRv6 Policy path quality of the first service does not meet the path quality requirement of the first service, selecting a Policy set meeting the path quality requirement from SRv6 policies corresponding to the first service; the Policy set contains at least one target Policy; reserving a first bandwidth for a first service according to each target Policy, and sending a deterministic probe message through the target Policy, so that when an intermediate node on a path of the target Policy determines that the idle bandwidth is not less than the first bandwidth, reserving the first bandwidth for the first service; judging whether target Policy which is reserved successfully by the intermediate nodes exists or not; if so, switching the first service to an entry Policy reserved successfully by the intermediate node. The method and the device can realize cooperative scheduling of the traffic among different head node devices, and avoid scheduling oscillation problems.

Description

Distributed flow scheduling method and device

Technical Field

The present invention relates to the field of communications technologies, and in particular, to a distributed traffic scheduling method and apparatus.

Background

In recent years, SRv (Segment Routing over Internet Protocol Version, IPv 6-based segment routing) financial backbone networks represented by four major lines have been built, and currently, four major lines each consider expanding SRv6 technology to branches/sites, and the whole network uses SRv6 technology, so an SDWAN (Software-Defined Wide Area Network ) SRv solution has been proposed.

In the SWDAN SRv6 scenario, traffic scheduling mainly uses a SPR (Smart Policy Route, intelligent policy routing) distributed traffic scheduling scheme.

In the SWDAN SRv6 distributed traffic scheduling scheme, the branch to headquarters establishes an end-to-end SRv Policy, and the path of each SRv6 Policy is not changed after being planned. In the flow scheduling process, a branch router executes a distributed intelligent routing strategy, and the branch router detects SRv end-to-end quality of the Policy by an IFIT (In-situ Flow Information Telemetry, flow detection), and the branch router automatically executes which SRv Policy the service walks.

However, in the existing distributed intelligent routing strategy, the scheduling results of different node devices are easy to collide, so that scheduling concurrency is caused.

Disclosure of Invention

The embodiment of the invention aims to provide a distributed flow scheduling method and device, so as to realize cooperative scheduling of flows among different head node devices and avoid scheduling oscillation. The specific technical scheme is as follows:

in a first aspect, an embodiment of the present invention provides a distributed traffic scheduling method, applied to a head node device in a SRv network, where the method includes:

when detecting that the path quality of the current IPv 6-based segment routing Policy SRv6 Policy of the first service does not meet the path quality requirement of the first service, selecting a Policy set meeting the path quality requirement from SRv6 policies corresponding to the first service; the Policy set comprises at least one target Policy;

reserving a first bandwidth for the first service according to each target Policy, and sending a deterministic probe message through the target Policy, so that when an intermediate node on a path of the target Policy determines that the idle bandwidth is not smaller than the first bandwidth, reserving the first bandwidth for the first service;

judging whether target Policy which is reserved successfully by the intermediate nodes exists or not;

if yes, switching the first service to an entry Policy which is reserved successfully by the intermediate node.

Optionally, the deterministic probe packet carries a deterministic check field, where the deterministic check field includes a service ID field, a source router ID field, a sink router ID field, a service bandwidth field, a retention time field, and/or a reserved status field;

the service ID field is used for indicating the type of the service;

the source router ID field is used for indicating a head node of the target Policy;

the sink router ID field is used for indicating the tail node of the target Policy;

the service bandwidth field is configured to indicate a bandwidth size reserved for a service by an intermediate node of the path of the target Policy;

the reserved time field is configured to instruct an intermediate node of the path of the target Policy to reserve a reserved time of a bandwidth for a service;

and the reserved state field is used for indicating whether the intermediate nodes of the path of the target Policy reserve bandwidth for the service indicated by the service ID field.

Optionally, the step of determining whether there is a target Policy that is reserved successfully by the intermediate node includes:

and receiving a reverse deterministic message fed back by the tail node of the target Policy after receiving the deterministic probe message, and judging whether the target Policy which is reserved successfully by the intermediate node exists or not based on a reserved state field in the reverse deterministic message.

Optionally, the first bandwidth filled in the service bandwidth field is determined according to a traffic statistic of the first service.

Optionally, the deterministic probe message is a flow-following detection IFIT message carrying the deterministic check field.

Optionally, the step of reserving a first bandwidth for the first service for each target Policy, and sending a deterministic probe packet through the target Policy, so that when determining that the idle bandwidth is not less than the first bandwidth, an intermediate node on a path of the target Policy reserves the first bandwidth for the first service includes:

and reserving the first bandwidth for the first service on an outgoing interface of the head node equipment in the path of the target Policy aiming at each target Policy, and sending a deterministic detection message through the target Policy, so that when the idle bandwidth is not less than the first bandwidth, an intermediate node on the path of the target Policy reserves the first bandwidth for the first service on the outgoing interface of the intermediate node in the path of the target Policy.

Optionally, the step of switching the first service to an intermediate node reserves a successful Policy entry, includes:

When detecting that the path quality of the current SRv Policy of one or more second services does not meet the path quality requirement of the second service, any target Policy of the second service selected from SRv6 policies corresponding to the second service is the same as one item of the first service;

the method further comprises the steps of: reserving a second bandwidth for the second traffic;

for the target Policy that the first service and the second service are the same, the step of sending a deterministic probe message through the target Policy, so that when determining that the idle bandwidth is not less than the first bandwidth, an intermediate node on a path of the target Policy reserves the first bandwidth for the first service includes:

transmitting a deterministic probe message through a target Policy with the same first service and second service, so that when an intermediate node on a path of the target Policy determines that an idle bandwidth is not smaller than the first bandwidth and/or the second bandwidth, the intermediate node reserves the first bandwidth for the first service and/or reserves the second bandwidth for the second service;

the method further comprises the steps of:

judging whether target polices which are reserved successfully by intermediate nodes exist in the target polices corresponding to the second service aiming at the second service; if yes, the second service is switched to the target Policy.

In a second aspect, an embodiment of the present invention provides a distributed traffic scheduling method, applied to an intermediate node device in a SRv network, including:

receiving a certainty detection message sent by a last hop node in a target Policy path of a first service; the target Policy is SRv Policy which is selected from SRv6 Policy corresponding to the first service and meets the path quality requirement when the path quality of the current SRv6 Policy of the first service does not meet the path quality requirement of the first service; the deterministic probe message comprises a reserved state field, and the reserved state field is used for indicating whether all the path nodes of the deterministic probe message in the path of the target Policy reserve a first bandwidth for the first service;

judging whether upstream nodes in the path of the target Policy reserve the first bandwidth for the first service or not based on a reserved state field in the deterministic probe message;

if not, forwarding the deterministic probe message to a next hop node in the path of the target Policy;

if yes, reserving the first bandwidth for the first service on an outgoing interface in the path of the target Policy under the condition that the idle bandwidth is not smaller than the first bandwidth, and forwarding the deterministic probe message to a next hop node in the path of the target Policy; and updating a reserved state field in the deterministic probe message under the condition that the idle bandwidth is smaller than the first bandwidth, and forwarding the updated deterministic probe message to a next hop node in the path of the target Policy.

Optionally, the method further comprises:

receiving a reverse deterministic message fed back by a tail node in the path of the target Policy after receiving the deterministic probe message, and judging whether intermediate nodes in the path of the target Policy reserve the first bandwidth for the first service or not based on a reserved state field in the reverse deterministic message;

if not, releasing the first bandwidth under the condition that the first bandwidth is reserved for the first service;

if yes, after the target Policy is switched for the first service, the first bandwidth reserved for the first service is released.

In a third aspect, an embodiment of the present invention provides a distributed traffic scheduling method, applied to a tail node device in a SRv network, including:

after receiving a determination detection message sent by a head node device in a target Policy path through the target Policy, adding a reverse identifier into the determination detection message to obtain a reverse determination message; the target Policy is SRv Policy which is selected from SRv6 Policy corresponding to the first service and meets the path quality requirement when the path quality of the current SRv6 Policy of the first service does not meet the path quality requirement of the first service;

And returning the reverse deterministic message through the path of the target Policy, wherein the reverse deterministic message is used for informing an upstream node in the path of the target Policy whether intermediate nodes in the path of the target Policy reserve a first bandwidth for a first service.

In a fourth aspect, an embodiment of the present invention provides a distributed traffic scheduling apparatus, applied to a head node device in a SRv network, including:

a selecting module, configured to select, when it is detected that a path quality of a current IPv 6-based segment routing Policy SRv Policy of a first service does not meet a path quality requirement of the first service, a Policy set that meets the path quality requirement from SRv policies corresponding to the first service; the Policy set comprises at least one target Policy;

the reservation module is used for reserving a first bandwidth for the first service according to each target Policy, and sending a deterministic detection message through the target Policy, so that when an intermediate node on a path of the target Policy determines that the idle bandwidth is not less than the first bandwidth, the first bandwidth is reserved for the first service;

the first judging module is used for judging whether target Policy which is reserved successfully by the intermediate node exists or not;

And the switching module is used for switching the first service to the one item of Policy which is reserved successfully in the intermediate node if the judging result of the first judging module is that the first service exists.

In a fifth aspect, an embodiment of the present invention provides a distributed traffic scheduling apparatus, applied to an intermediate node device in a SRv network, including:

the receiving module is used for receiving a certainty detection message sent by a last hop node in a target Policy path of the first service; the target Policy is SRv Policy which is selected from SRv6 Policy corresponding to the first service and meets the path quality requirement when the path quality of the current SRv6 Policy of the first service does not meet the path quality requirement of the first service; the deterministic probe message comprises a reserved state field, and the reserved state field is used for indicating whether all the path nodes of the deterministic probe message in the path of the target Policy reserve a first bandwidth for the first service;

the second judging module is used for judging whether the upstream nodes in the path of the target Policy reserve the first bandwidth for the first service or not based on the reserved state field in the deterministic probe message;

The forwarding module is used for forwarding the deterministic detection message to a next hop node in the path of the target Policy if the judgment result of the second judging module is negative; if the judgment result of the second judgment module is yes, reserving the first bandwidth for the first service on an output interface in the path of the target Policy under the condition that the idle bandwidth is not smaller than the first bandwidth, and forwarding the deterministic probe message to a next hop node in the path of the target Policy; and updating a reserved state field in the deterministic probe message under the condition that the idle bandwidth is smaller than the first bandwidth, and forwarding the updated deterministic probe message to a next hop node in the path of the target Policy.

In a sixth aspect, an embodiment of the present invention provides a distributed traffic scheduling apparatus, applied to a tail node device in a SRv network, including:

the adding module is used for adding a reverse identifier into the deterministic detection message after receiving a deterministic detection message sent by the head node equipment in the target Policy path through the target Policy, so as to obtain a reverse deterministic message; the target Policy is SRv Policy which is selected from SRv6 Policy corresponding to the first service and meets the path quality requirement when the path quality of the current SRv6 Policy of the first service does not meet the path quality requirement of the first service;

And the return module is used for returning the reverse deterministic message through the path of the target Policy, and the reverse deterministic message is used for informing an upstream node in the path of the target Policy whether intermediate nodes in the path of the target Policy reserve a first bandwidth for the first service.

In a seventh aspect, an embodiment of the present invention provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, where the processor, the communication interface, and the memory complete communication with each other through the communication bus;

a memory for storing a computer program;

and the processor is used for realizing any distributed flow scheduling method in the first aspect when executing the program stored in the memory.

The embodiments of the present invention also provide a computer program product comprising instructions which, when run on a computer, cause the computer to perform any of the above-described distributed traffic scheduling methods.

The embodiment of the invention has the beneficial effects that:

according to the distributed traffic scheduling method and device provided by the embodiment of the invention, when the head node equipment detects that the current SRv Policy path quality of a first service does not meet the path quality requirement of the first service, a Policy set meeting the path quality requirement is selected from SRv policies corresponding to the first service, a first bandwidth is reserved for the first service for each target Policy before the first service is switched, a determination detection message is sent through the target Policy, so that intermediate nodes on the path of the target Policy reserve the first bandwidth for the first service when the idle bandwidth is not less than the first bandwidth, then whether the target Policy which is reserved successfully by the intermediate nodes exists is judged, and when the first service is switched, the first service is switched to the first item of the target Policy which is reserved successfully by the intermediate nodes.

Under the condition that a plurality of head node devices in the network all need to schedule the traffic of the service on the head node, the intermediate node reserves the bandwidth for the service on the corresponding head node device after receiving the deterministic probe message sent by the head node device, so the bandwidth reserved by the intermediate node for the service on one head node device cannot be reserved for the service on other head node devices. And under the condition that the target Policy which is successfully reserved by the intermediate node exists in the target Policy determined by the first node equipment, the first service on the first node equipment is switched to the target Policy which is successfully reserved by one of the intermediate nodes. Therefore, in the process of scheduling the service flow, the plurality of head node devices cannot preempt the bandwidth on the intermediate node, so that the cooperative scheduling of the service flow among different head node devices is indirectly realized, the problem of SRv Policy path quality degradation, invalid scheduling or error scheduling of the service flow and scheduling oscillation caused by the fact that the service on different head node devices is directly switched to a new SRv Policy and then preempted the bandwidth is avoided under the condition that the service flow is scheduled in an isolated mode by each head node device is avoided, and the certainty of distributed flow scheduling is improved.

Of course, it is not necessary for any one product or method of practicing the invention to achieve all of the advantages set forth above at the same time.

Drawings

In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other embodiments may be obtained according to these drawings to those skilled in the art.

FIG. 1 is a schematic diagram of a network architecture of a financial networking;

FIG. 2 is a schematic diagram of an implementation mechanism of a distributed intelligent routing policy;

FIG. 3 is a schematic path diagram of a distributed traffic scheduling scheme;

fig. 4 is a schematic flow chart of a distributed traffic scheduling method according to an embodiment of the present invention;

fig. 5 is another flow chart of a distributed traffic scheduling method according to an embodiment of the present invention;

fig. 6 is a schematic flow chart of a distributed traffic scheduling method according to an embodiment of the present invention;

fig. 7 is a schematic structural diagram of a distributed flow scheduling device according to an embodiment of the present invention;

fig. 8 is another schematic structural diagram of a distributed flow scheduling device according to an embodiment of the present invention;

Fig. 9 is a schematic structural diagram of a distributed flow scheduling device according to an embodiment of the present application;

fig. 10 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

The following description of the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present application, but not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by the person skilled in the art based on the present application are included in the scope of protection of the present application.

In recent years, SRv financial backbone networks represented by four major activities have been constructed, and currently, the four major activities consider expanding SRv technology to branches/sites, and the whole network uses SRv technology, so that an SDWAN SRv6 scheme is proposed.

Because the SDWAN SRv6 expands the networking to branches/mesh points, the traditional centralized wide area backbone traffic scheduling scheme is not suitable for the SDWAN scenario, and the SPR distributed traffic scheduling scheme is considered to be used in this scenario, mainly for the following reasons:

1. the conventional wide area backbone traffic scheduling scheme is to separately stream each service to one SRv Policy, or SR Policy, and each service separately performs end-to-end traffic scheduling based on SRv path change of the Policy. However, in the SDWAN scenario, the number of branches/mesh points may reach tens of thousands, which is far greater than the number of access points in the wide area network, and meanwhile, the service classification is more than that in the backbone network. Therefore, if traffic is scheduled in the SDWAN scenario according to the conventional wide area network scheduling scheme, the number of policies SRv, 6, may be very large, and the node devices within the network may not support the specifications of the policies.

2. The branches/sites are low-end devices, the performance is lower, and the number of supportable SRv Policy is less than that of wide-area backbone network node devices.

3. The increase of SRv Policy number can lead to higher pressure when the controller performs centralized scheduling, scheduling may not be timely, and the timeliness of scheduling can be guaranteed by the distributed flow scheduling scheme.

4. The characteristics of the financial networking determine that the number of paths of the branch access network is not as large as that of the wide area network, and the total possible paths are controllable.

Fig. 1 is a schematic diagram of a network structure of a financial networking, and an exemplary application scenario of the distributed traffic scheduling method provided in the embodiment of the present invention is described below with reference to fig. 1.

The SDWAN formed by a general line, a branch line and a plurality of branches/nodes is specifically shown in fig. 1, where the general line and the branch line are interconnected through a wide area backbone network, and the branches/nodes are accessed to the wide area backbone network through an access backbone network. The end-to-end SRv Policy is built from branches/dots to the headquarters, generally, each branch/dot to the headquarters has 2-3 end-to-end SRv policies, and paths of these SRv policies are not changed after being planned.

In the process of carrying out service flow scheduling, a branch router executes a distributed intelligent routing strategy, the branch router detects the end-to-end path quality of each SRv Policy in real time through IFIT, and a branch routing device automatically executes which SRv Policy the service walks.

Fig. 2 is a schematic diagram of an implementation mechanism of the distributed intelligent routing policy, and for convenience of understanding, one possible implementation mechanism of the distributed intelligent routing policy is described below by way of example with reference to fig. 2.

The specific structure of SRv6 Policy will be described first.

In the SRv network, each SRv6 Policy can be uniquely identified by a head node, a tail node, and a Color value, where the Color value is specifically an extended community attribute carried by the SRv6 Policy.

Specifically, each SRv Policy may correspond to multiple Candidate paths, where the Candidate paths are divided into Active paths and Standby paths, and different signaling protocols may issue different Candidate paths. Each Candidate Path is configured with a Segment List (Segment List), where the Segment List is used to identify a specific forwarding Path, for example, the Segment List may be "PE1- > P2- > PE3", and the token packet is sent to P2 by PE1 and then forwarded to PE3 by P2.

Also, SR Policy groups (SR Policy groups) may be configured in a SRv network, each SR Policy Group may be understood as a set of SRv6 policies with the same tail node, and these SRv6 policies may have different Color values.

When the service flow needs to be forwarded, the service flow is iterated to SR Policy groups through Color value drainage, and each SR Policy Group can be matched with a plurality of services.

Specifically, color values required by the service traffic may be preconfigured, and the service traffic may be drained to an SR Policy Group having a corresponding Color value according to the Color values required by the service traffic.

As an example, if traffic can be forwarded by SRv Policy with Color value 1, 2 or 3, and SR Policy Group 1 includes SRv Policy with Color value 1, 2 or 3, traffic can be drained to SR Policy Group 1.

And each service can be configured with an intelligent routing strategy, and the intelligent routing strategy can specifically comprise an end-to-end path quality requirement of the service and a group of SR Policy colors. The path quality requirements may include, among other things, various path switching metrics such as delay, jitter, and packet loss rate. A set of SR Policy Color may specifically include a plurality of Color values, SRv Policy with corresponding Color values may be used to forward traffic for the traffic, and one priority may be specified for each Color value therein, e.g., a set of SR Policy Color may include Color 1, color 2, and Color 3, where Color 1 has the highest priority, color 2 has the highest priority, and Color 3 has the lowest priority.

Referring to fig. 2, traffic flows led to the SR Policy Group include two traffic classes (Flow classes) of production traffic and office traffic, each of which is configured with an intelligent routing Policy.

According to the priority of the Color value configured in the intelligent routing Policy and the Color value of each SR Policy, the priority order of the plurality of SR policies for specific services can be determined. For the production service, SR Policy 1 has the highest priority (high priority), and SR Policy 2 has a lower priority (lower priority); for office services, SR Policy 2 has the highest priority and SR Policy 1 has a lower priority.

For the service with the configured intelligent routing Policy, the intelligent routing can perform IFIT detection on each SR Policy, obtain the path quality of a plurality of forwarding paths of forwarding path 1-forwarding path N based on IFIT detection data, and then select the path of the SR Policy with the highest priority and meeting the path quality requirement of the service from the SR Policy as the traffic forwarding path of the service.

For example, for the production service, SR Policy 1 has the highest priority, and if the path quality of SR Policy 1 meets the path quality requirement of the production service, the traffic of the production service is drained to SR Policy 1. And (3) guiding the service to an SR Policy, namely forwarding the traffic of the service through the SR Policy.

After traffic is drained to SRv Policy with high priority, if SRv Policy with high priority is invalid or path quality is degraded, the traffic can be switched to SRv6 Policy with low priority, and if path quality of SRv6 Policy with high priority is recovered during the period of taking SRv Policy with low priority as a traffic forwarding path, the traffic can be switched back to SRv Policy with high priority.

To avoid scheduling concurrency, the switching and back-off times of traffic between SRv policies may also be set.

Fig. 3 is a schematic path diagram of a distributed traffic scheduling scheme, and a specific procedure for scheduling traffic in the distributed traffic scheduling scheme is exemplarily described with reference to fig. 3.

Referring to fig. 3, fig. 3 specifically shows two head nodes of PE1 and PE3, where service 1 (Flow Class 1) and service 2 (Flow Class 2) exist on both PE1 and PE3, bandwidths of service 1 and service 2 are 0.8G, service 1 and service 2 are both introduced into tail node as PE2, color values include SR Policy Group 1 of 1, 2 and 3, and SR Policy Group 1 is configured on both PE1 and PE 3.

The service 1 is configured with an intelligent routing strategy SPR1, the path quality requirement is that the delay is not more than 100ms, the jitter is not more than 30ms, the packet loss rate (loss) is not more than 1%, and the Color value priority order corresponding to the service 1 is Color 1,Color 2,Color 3 according to the priority order.

The service 2 is configured with intelligent route measurement SPR2, the path quality requirement is that the delay is not more than 200ms, the jitter is not more than 60ms, the packet loss rate is not more than 2%, and the priority order of Color values corresponding to the service 2 is Color 3, color 2 and Color 1 according to the priority order from high to low.

The first node PE1 is configured with three SRv6 policies, namely SR Policy PE1_1,SR Policy PE1_2 and SR Policy PE1_3, wherein the Color value of SR Policy PE1_1 is 1, the first node is PE1, the tail node is PE2, and the segment list is PE1- > P1- > PE2; the Color value of SR Policy PE 1-2 is 2, the first node is PE1, the tail node is PE2, and the segment list is PE1- > P2- > PE2; the Color value of SR Policy PE 1-3 is 3, the first node is PE1, the tail node is PE2, and the segment list is PE1- > P3- > PE2. The SR Policy PE 3-1,SR Policy PE3_2 and SR Policy PE 3-3 are configured on the PE3 to form three SRv6 policies, wherein the Color value of the SR Policy PE 3-1 is 1, the first node is PE3, the tail node is PE2, and the segment list is PE3- > P1- > PE2; the Color value of SR Policy PE 3-2 is 2, the first node is PE3, the tail node is PE2, and the segment list is PE3- > P2- > PE2; the Color value of SR Policy PE 3-3 is 3, the first node is PE3, the tail node is PE3, and the segment list is PE3- > P3- > PE2.

In addition, in the network shown in fig. 3, three interfaces of each node are illustrated and labeled G1/0, G2/0, and G3/0, respectively, the bandwidths of the links from the head node devices PE1, PE3 to the intermediate nodes P1, P2, P3 are all 1G, the bandwidth of the link between P2 and PE2 is 1G, and the bandwidth of the link between P1 and PE2 and between P3 and PE2 is all 2.5G.

In addition, each SRv Policy may initiate an IFIT probe, so that each of the head nodes PE1 and PE3 may obtain the path quality of each SRv Policy configured at the node.

When initially scheduling traffic of traffic 1 and traffic 2, color 1 has the highest priority for traffic 1, so traffic of traffic 1 on PE1 and PE3 is forwarded through SR Policy PE1_1 and SR Policy PE3_1 of Color 1, i.e. traffic of traffic 1 on PE1 is forwarded through the path of PE1- > P1- > PE2, and traffic of traffic 1 on PE2 is forwarded through the path of PE3- > P1- > PE 2. Similarly, color 3 has the highest priority for service 2, and traffic of service 2 on PE1 and PE3 is forwarded through SR Policy PE1_3 and SR Policy PE3_3 of Color 3, i.e. traffic of service 2 on PE1 is forwarded through the path of PE1- > P3- > PE2, and traffic of service 2 on PE2 is forwarded through the path of PE3- > P3- > PE 2.

When the link from P3 to PE2 fails or the quality deteriorates, the first nodes PE1 and PE3 detect that the path quality of SR Policy pe1_3 and SR Policy pe3_3 of Color 3 do not meet the path quality requirement of service 2, and Color 2 has the next highest priority for service 2, so PE1 and PE3 perform distributed scheduling on service 2 at the same time, and schedule the traffic of service 2 to SR Policy pe1_2 and SR Policy pe3_2 of Color 2, respectively.

In the foregoing, the bandwidth of the service 2 on the PE1 and the PE3 is 0.8G, and the SR Policy pe1_2 and the SR Policy pe3_2 occupy the link from P2 to PE2, and the bandwidth of the link is only 1G, so after the PE1 and the PE3 schedule the traffic of the service 2 to the SR Policy pe1_2 and the SR Policy pe3_2 respectively, congestion and packet loss may occur in the link from P2 to PE 2.

Under the condition that congestion and packet loss occur in a link from P2 to PE2, PE1 and PE3 respectively detect that the path quality of SR Policy PE1_2 and SR Policy PE3_2 do not meet the path quality requirement of service 2, PE1 and PE3 can perform distributed scheduling on service 2 again at the same time, and respectively switch the traffic of service 2 to SR Policy PE1_1 and SR Policy PE3_1 of Color 1.

In the foregoing, the traffic of the service 1 on the PE1 and the PE2 is also forwarded through the SR Policy PE1_1 and the SR Policy PE3_1, respectively, and both the SR Policy PE1_1 and the SR Policy PE3_1 occupy the bandwidth of the link from P1 to PE2, which easily causes congestion and packet loss on the link from P1 to PE 2. At this time, PE1 and PE3 respectively detect that the path quality of SR Policy PE1_1 and SR Policy PE3_1 do not meet the path quality requirements of service 1 and service 2, and PE1 and PE3 simultaneously perform distributed scheduling again, so as to switch the traffic of part of the services to SR Policy PE1_2 and SR Policy PE3_2 of Color 2 simultaneously. Therefore, the head nodes respectively perform independent scheduling, and cannot realize cooperative scheduling, so that frequent scheduling and invalid scheduling are caused, and scheduling oscillation is generated.

In order to solve the problem of scheduling oscillation generated in the process of distributed traffic scheduling, the embodiment of the invention provides a distributed traffic scheduling method, which is specifically applied to a first node device in a SRv network, and fig. 4 is a schematic flow chart of the distributed traffic scheduling method provided in the embodiment of the invention, and referring to fig. 4, the method specifically includes the following steps:

step S401: when detecting that the current SRv6 Policy path quality of the first service does not meet the path quality requirement of the first service, selecting a Policy set meeting the path quality requirement from SRv6 policies corresponding to the first service; the Policy set contains at least one target Policy.

The current SRv6 Policy of the first traffic, namely SRv6 Policy currently used to forward traffic of the first traffic. As an example, in practical applications, different services in a network may be distinguished by a service ID (IDentity, identification number).

Specifically, for each type of service, a corresponding path quality requirement can be configured, and the specific content of the path quality requirement can be configured according to actual requirements. By way of example, path quality requirements may include requirements for delay, jitter, and packet loss rate, and reference may be made to the foregoing description of intelligent routing strategies for specific applications of path quality requirements.

The plurality of SRv6 policies are specifically configured on the head node device in the SRv network, and during the process of scheduling the traffic, the head node device in the SRv6 network can detect the path quality according to the SRv6 policies configured on the node, and determine SRv whether the path quality of the 6 policies meets the path quality requirement of the currently forwarded traffic. As one example, the head-node device may obtain SRv Policy's end-to-end path quality through IFIT detection.

When detecting that the current SRv6 Policy path quality of any service does not meet the path quality requirement of the service, the head node device can schedule the service. The method for distributed traffic scheduling according to the embodiment of the present invention is described below by taking the first service as an example.

The SRv Policy corresponding to the first service may be understood as SRv6 Policy configured in SRv Policy on the head node device, which may be used to transmit the traffic of the first service.

Specifically, for a first service on the head node, whose tail node in the SRv network is determined, each SRv Policy in the SRv6 network has one Color value, each service may correspond to one or more Color values, and traffic for the service may be transmitted through the SRv6 policies with those Color values. Therefore, SRv6 Policy corresponding to the first service can be understood as SRv6 Policy having the same tail node as the current SRv6 Policy and Color value corresponding to the first service.

When detecting that the current SRv6 Policy of the first service does not meet the path quality requirement of the first service, one or more SRv6 policies meeting the path quality requirement of the first service can be selected from SRv6 policies corresponding to the first service, and a Policy set is obtained.

Alternatively, it may be understood that the first service is already introduced into the SRv Policy Goup with the same tail node and meeting the Color value requirement of the first service when the first service performs initial scheduling, and when it is detected that the current SRv6 Policy of the first service does not meet the path quality requirement of the first service, a Policy set meeting the path quality requirement of the first service is selected from the SRv6 Policy Goup.

Step S402: and reserving a first bandwidth for the first service according to each target Policy, and sending a deterministic probe message through the target Policy, so that an intermediate node on a path of the target Policy reserves the first bandwidth for the first service when determining that the idle bandwidth is not less than the first bandwidth.

The Policy set determined in the foregoing step S401 may specifically include at least one target Policy.

For each target Policy, the head node device reserves bandwidth for the first service, sends a deterministic probe message through the target Policy, and reserves the first bandwidth for the first service when the idle bandwidth of the intermediate node is not less than the first bandwidth after the intermediate node on the path of the target Policy receives the deterministic probe message.

After the first node device and the intermediate node reserve the first bandwidth for the first service, the reserved bandwidth cannot be occupied by other services, and cannot be reserved for other services.

The first bandwidth may be understood as an estimated value of a bandwidth size required to be occupied when the traffic of the first service is forwarded through the target Policy, and the specific value may be configured according to an actual requirement.

In practical applications, it may be detected that the path quality of the current SRv Policy of the plurality of services does not meet the path quality requirement of the service, and therefore, when the first bandwidth is reserved for the first service, the first node device or the intermediate node essentially reserves the first bandwidth for the first service in the idle bandwidths that are not currently occupied and reserved for other services.

Specifically, an intermediate node in the SRv network may receive multiple deterministic probe messages, in which case the intermediate node may reserve bandwidth according to the order of receipt of the deterministic probe messages.

As an example, if P2 shown in fig. 3 receives a deterministic probe packet sent by PE1 through SR Policy PE1_2 first, and then receives a deterministic probe packet sent by PE3 through SR Policy PE3_2, P2 processes the deterministic probe packet sent by PE1 first, determines whether bandwidth can be reserved for a service on PE1, reserves a bandwidth for the service when the idle bandwidth is not less than the bandwidth required by the service, and then processes the deterministic probe packet sent by PE3, and determines whether bandwidth can be reserved for the service on PE3 according to the idle bandwidth.

Step S403: judging whether target Policy which is reserved successfully by the intermediate nodes exists or not;

if so, switching the first service to an entry Policy reserved successfully by the intermediate node.

After a deterministic probe message is sent through a target Policy, the head node device needs to determine whether all intermediate nodes on the path of the target Policy are reserved successfully, that is, the first bandwidth is reserved for the first service, and if all intermediate nodes reserve the first bandwidth for the first service, the target Policy can meet the bandwidth requirement of the first service.

If any intermediate node on a target Policy path fails to reserve, i.e. the first bandwidth is not reserved for the first service, the idle bandwidth of any intermediate node may be smaller than the first bandwidth, and the target Policy cannot meet the bandwidth requirement of the first service.

Therefore, the first node device can specifically switch the first service to an entry target Policy that is reserved successfully by the intermediate node, and can specifically select one entry target Policy according to actual requirements to switch the first service to the target Policy when the intermediate node with a plurality of target policies is reserved successfully by the intermediate node. And switching the first service to a target Policy, specifically switching the flow of the first service to the target Policy.

In addition, when the head node device needs to send the deterministic probe message through the plurality of target polices, a mechanism and an order for sending the deterministic probe message through each target Policy can be selected according to actual requirements, and the embodiment of the invention is not limited to this.

As an example, the head node device may reserve the first bandwidth for the first service for multiple target polices at the same time, and send the deterministic probe packet through these target polices at the same time, so that the intermediate node on the paths of the multiple target polices reserves the bandwidth for the first service, and thus the head node device may determine, with higher efficiency, the target polices that may be used to carry the first service.

As another example, the head node device may also send the deterministic probe packet sequentially through each target Policy, and send the deterministic probe packet through the next target Policy when reservation of the intermediate node on the current target Policy path fails, and directly switch the first service to the target Policy when reservation of the intermediate node on the current target Policy path is successful, so as to save bandwidth occupation in the SRv network.

Taking fig. 3 as an example, when the head node device PE1 detects that SR Policy PE1_3 does not meet the path quality requirement of service 2, and SR Policy PE1_1 and SR Policy PE1_2 corresponding to service 2 meet the path quality requirement of service 2, PE1 may send a deterministic probe packet through SR Policy PE1_1 and SR Policy PE1_2 at the same time, so that intermediate nodes P1 and P2 reserve bandwidth for service 2, and then determine whether intermediate nodes on paths of SR Policy PE1_1 and SR Policy PE1_2 are reserved successfully, if SRv Policy which is reserved successfully by intermediate nodes exists, one intermediate node in service 2 is reserved successfully SRv6 Policy. The PE1 may also send a deterministic probe packet through SR Policy PE1_2 first, if the reservation of the intermediate node P2 on the path of SR Policy PE1_2 is successful, then directly switch the service 2 to SR Policy PE1_2, if the reservation of the intermediate node P2 fails, then send a deterministic probe packet through SR Policy PE1_1 to make the intermediate node P1 on the path of SR Policy PE1_1 reserve bandwidth for the service 2, if the reservation of the intermediate node P1 is successful, then switch the service 2 to SR Policy PE1_1, and if the reservation of the intermediate node P1 fails, then determine that the routing of the service 2 fails.

In the embodiment of the invention, the specific mode of judging whether the intermediate nodes on the path of any target Policy reserve the first bandwidth for the first service by the head node equipment is not limited. As an example, an analyzer may be deployed in the network, and the intermediate node may report the reservation result to the analyzer after determining whether the node is capable of reserving the first bandwidth for the first service, where the analyzer analyzes whether the intermediate nodes on the target Policy path are all reserved successfully based on the reservation result reported by each intermediate node on the path of the target Policy, and issues the analysis result to the head node device.

In the embodiment of the present invention, the at least one target Policy determined in the foregoing step S401 may be understood as SRv Policy that is considered to be used for forwarding the traffic of the first service after performing the preliminary analysis based on the path quality requirement of the first service and the path quality of SRv Policy.

Before switching the traffic of the first service to a target Policy, the head node device reserves a first bandwidth for the first service, and sends a deterministic probe message through the target Policy, so that an intermediate node on the path of the target Policy reserves the first bandwidth for the first service, and when determining that there is a target Policy that the intermediate node reserves successfully, the traffic of the first service is switched to a target Policy that the intermediate node reserves successfully.

The process that the head node device reserves a first bandwidth for the first service and sends the deterministic probe message through the target Policy can be understood as performing deterministic check on the target Policy. The process can specifically determine whether intermediate nodes on the path of the target Policy have sufficient bandwidth remaining for transmission of traffic of the first service, and reserve bandwidth required for forwarding traffic of the first service on the head-node device and the intermediate nodes.

Under the condition that the first node equipment and the intermediate node in the path of the target Policy reserve a first bandwidth for the first service, namely the certainty check of the target Policy is successful, the path of the target Policy is considered to be capable of meeting the bandwidth requirement of the first service. Under the condition that any intermediate node in the path of the target Policy does not reserve the first bandwidth for the first service, namely the certainty check of the target Policy fails, the path of the target Policy is considered to be unable to meet the bandwidth requirement of the first service, so that the first service can be switched to the one item of target Policy which is reserved successfully by the intermediate node.

In the following, referring to fig. 3, it is assumed that, in the initial case, service 1 on PE1 is forwarded through SR Policy PE1_1, service 2 is forwarded through SR Policy PE1_3, service 1 on PE3 is forwarded through SR Policy PE3_1, and service 2 is forwarded through SR Policy PE 3_3.

When the link from P3 to PE2 fails or the path quality deteriorates, PE1 and PE3 detect that the path quality of SR Policy pe1_3 and SR Policy pe3_3, respectively, do not meet the path quality requirement of service 2, and PE1 determines SR Policy pe1_2 as the target Policy of service 2, and PE2 determines SR Policy pe3_2 as the target Policy of service 2. If the intermediate node P2 remains 1G bandwidth and reserves 0.8G bandwidth for the service 2 on the PE1, then P2 only remains 0.2G bandwidth, and cannot reserve 0.8G bandwidth for the service 2 on the PE2, so that only the PE1 can switch the traffic of the service 2 to SR Policy PE1_2, and the PE3 cannot switch the traffic of the service 2 to SR Policy PE2_2, so that the P1 and PE3 cannot preempt the bandwidth on the P2, and further congestion and packet loss occur in the link from the P2 to the PE2, thereby avoiding the scheduling oscillation problem.

According to the distributed traffic scheduling method provided by the embodiment of the invention, when the head node equipment detects that the current SRv Policy path quality of a first service does not meet the path quality requirement of the first service, a Policy set meeting the path quality requirement is selected from SRv policies corresponding to the first service, a first bandwidth is reserved for the first service for each target Policy before the first service is switched, and a determination detection message is sent through the target Policy, so that when an intermediate node on the path of the target Policy is determined to be not smaller than the first bandwidth, the first bandwidth is reserved for the first service, then whether the target Policy which is reserved successfully by the intermediate node exists is judged, and when the first service is switched, the first service is switched to the first item of Policy which is reserved successfully by the intermediate node.

In one embodiment of the present invention, the step S402 specifically includes:

and reserving a first bandwidth for a first service on an output interface of a head node device in a path of each target Policy according to each target Policy, and sending a deterministic probe message through the target Policy, so that when an intermediate node on the path of the target Policy determines that the idle bandwidth is not less than the first bandwidth, the first bandwidth is reserved for the first service on the output interface of the intermediate node in the path of the target Policy.

Specifically, each head node device or intermediate node may have multiple interfaces thereon, and in the embodiment of the present invention, when the head node device or intermediate node reserves a bandwidth, a first bandwidth is specifically reserved for a first service on an outgoing interface in a path of the target Policy.

For the exemplary illustration in connection with fig. 3, assume that PE1 determines SR Policy PE1_2 as an entry for service 2, the path of the target Policy is PE1- > P2- > PE2. The first node device PE1 specifically has three interfaces G1/0, G2/0 and G3/0, and the PE1 is connected to the next hop node P2 through the outgoing interface G2/0, so that the PE1 reserves a first bandwidth for the first service on the outgoing interface G2/0, the intermediate node P2 specifically has three interfaces G1/0, G2/0 and G3/0, and the P2 is connected to the next hop node PE2 through the outgoing interface G3/0, so that the P2 reserves a first bandwidth for the first service on the outgoing interface G3/0.

In practical application, after the flow of the first service is switched to an entry Policy, the flow of the first service specifically occupies the bandwidth of an outgoing interface of the first node device in the path of the target Policy and the bandwidth of an outgoing interface of the intermediate node in the path of the target Policy, so that the first bandwidth is reserved for the first service on the outgoing interface of the first node device in the path of the target Policy, the first bandwidth is reserved for the first service on the outgoing interface of the intermediate node in the path of the target Policy, the practical application requirement is met, whether the flow of the first service can be switched to the target Policy or not is judged based on the actual application requirement, the accuracy is higher, and the effect of avoiding scheduling oscillation is better.

In addition, before sending the deterministic probe message through an entry Policy, the head node device may specifically further determine whether the idle bandwidth on the output interface in the target Policy path is not less than the first bandwidth, if yes, send the deterministic probe message through the target Policy, and if no, directly determine that the deterministic check of the target Policy fails, and no more send the deterministic probe message through the target Policy.

Taking the situation that the head node equipment sequentially performs deterministic verification on different target polices as an example, if the idle bandwidth of the head node equipment on an output interface in a current target Policy path is smaller than a first bandwidth, determining that the deterministic verification of the current target Policy fails, selecting another SRv6 Policy from a Policy set corresponding to the determined first service as a new target Policy, and performing deterministic verification on the target Policy, thereby saving time for performing deterministic verification on the target Policy and improving efficiency of distributed scheduling of flows.

Taking fig. 3 as an example, assume that PE1 determines SR Policy pe1_2 as a target Policy of service 2, PE1 first determines whether bandwidth can be reserved for service 2 on an outgoing interface G2/0 in a path of SR Policy pe1_2, and if bandwidth cannot be reserved for service 2 on an outgoing interface G2/0, a deterministic check message is not sent through SR Policy pe1_2 any more, but SR Policy pe1_3 can be directly determined as a new target Policy, and deterministic check is performed on SR Policy pe1_3.

In one embodiment of the present invention, the deterministic probe message carries a deterministic check field, where the deterministic check field includes a traffic ID field, a source router ID field, a sink router ID field, a traffic bandwidth field, a retention time field, and/or a reservation status field;

the service ID field is used for indicating the type of the service;

a source router ID field for indicating a head node of the target Policy;

a sink router ID field for indicating a tail node of the target Policy;

a service bandwidth field, configured to indicate a bandwidth size reserved for a service by an intermediate node of a path of the target Policy;

a retention time field, configured to indicate a retention time of reserving a bandwidth for a service by an intermediate node of a path of the target Policy;

A reserved status field, configured to indicate whether intermediate nodes of the path of the target Policy reserve bandwidth for the traffic indicated by the traffic ID field.

In practical application, multiple service flows may enter an ingress interface of a head node device, and in the embodiment of the present invention, it is necessary to distinguish the flows of different services, and identify the flows of different services through service IDs filled in a service ID field, where specific content of the service ID may be selected according to practical requirements.

Specifically, a matching rule can be preset on the head node equipment, and the characteristics of the traffic on the inlet interface of the head node equipment are matched based on the matching rule, so that the traffic of different services is identified.

As an example, the matching rule may be an L3 ACL (Layer 3 Access Control List,L3 Layer access control list), an application group, or an L2 (Layer 2) type. Wherein the L3 ACL specifically includes a source IP (Internet Protocol ) address, a destination IP address, a source port number, a destination port number, a protocol type, and DSCP (Differentiated Services Code Point ); the application group is realized by deep message identification matching; the L2 types include specifically a source MAC (Media Access Control, medium access control), a destination MAC, L2 protocol, 8021P (a traffic priority control standard) and VLAN ID (Virtual Local Area Network ID ).

The service ID field may be, for example, an integer type.

The source router ID field may be, for example, an IPv6-address type, in which the IPv6 address of the active router, i.e., the unique identification of the source router in the IPv6 network, is filled.

Illustratively, the sink router ID field may specifically be of the IPv6-address type, populated with the IPv6 address of the sink router, i.e., the unique identity of the sink router in the IPv6 network.

The service bandwidth size filled in the service bandwidth field can also be understood as an estimated value of the bandwidth size required to be occupied in the link of the target Policy after the traffic of the service is switched to the target Policy, and the estimated value can be obtained through measurement.

The traffic bandwidth field may be specifically a Long (one type of data) type, in units of bps (bits per second).

Illustratively, the retention time field may be of integer type, in s (seconds), in particular.

The length of the retention time may be defined according to actual requirements, for example, in the case of sequentially performing deterministic verification on each target Policy, the retention time may be set to be not less than a duration of a detection period x a preferred path number, where the detection period may be understood as a duration required for deterministic detection on one target Policy, and the preferred path number is a total number of SRv6 policies on the head node device that has the same tail node as the current SRv Policy of the first service and meets a path quality requirement of the first service.

Specifically, the time during which the intermediate node reserves the first bandwidth for the first traffic exceeds the reservation time, and this portion of the first bandwidth will be released.

For example, the intermediate nodes of the path that can fill 1 in the reservation status field to characterize the target Policy reserve bandwidth for traffic, i.e., in the bandwidth reservation, any intermediate node of the path that fills 0 in the reservation status field to characterize the target Policy does not reserve bandwidth for traffic, i.e., the bandwidth reservation fails.

Any intermediate node on the path of the target Policy receives the deterministic probe message sent by the last hop node, after judging whether the node can reserve the first bandwidth for the first service, updates the reserved state field in the deterministic probe message according to the judging result, and then forwards the deterministic probe message to the next hop node.

Taking fig. 3 as an example, if PE1 determines SR Policy PE1_2 as a target Policy of service 2 and reserves bandwidth for service 2, then a deterministic probe packet is sent through SR Policy PE1_2, a reserved status field in the deterministic probe packet indicates that reservation is successful, after receiving the deterministic probe packet, P2 updates a reserved status field in the deterministic probe packet if bandwidth cannot be reserved for service 2, and forwards the updated deterministic probe packet to PE2, where the reserved status field in the deterministic probe packet indicates that reservation is failed.

As an example, the intermediate node on the path of the target Policy may update the reserved bandwidth data table recorded in the local content after receiving the deterministic probe message, and the reserved bandwidth data table of the intermediate node P2 shown in fig. 3 may be as shown in table 1:

TABLE 1

Referring to table 1, it can be seen that the reserved bandwidth data table specifically records the out interface name, the source router ID, the sink router ID, the service bandwidth size, the reserved time, and the reserved state.

Referring to fig. 3, if PE1 determines SR Policy pe1_2 as the target Policy of service 2, PE3 determines SR Policy pe3_2 as the target Policy of service 2, the deterministic probe packet sent by PE1 is transmitted through path PE1- > P2- > PE2, and the deterministic probe packet sent by PE3 is transmitted through path PE3- > P2- > PE2. If the deterministic probe message sent by the PE1 arrives at P2 first, P2 can reserve a bandwidth of 0.8G for the service 2 on the egress interface G3/0. Therefore, the information shown in the second row in the table 1 is recorded in the reserved bandwidth data table, specifically, the service with the service ID of 2 on the outgoing interface G3/0 is characterized in that the service 2 reserves the bandwidth of 800000000bps, the source router ID of the traffic of the service is 1:1, that is, the IPv6 address of PE1, the destination router ID is 2:1, that is, the IPv6 address of PE2, that is, the first node of the current SRv Policy of the service is PE1, and the tail node is PE2. And P2 reserves the bandwidth for service 2 for a reservation time of 3s, and the reserved state of the part of bandwidth is reserved.

After P2 receives the deterministic probe message sent by PE1, P2 receives the deterministic probe message sent by PE3 again, so that the contents shown in the third row in table 1 are recorded in the reserved bandwidth data table, and the specific meaning can be referred to in the description above. Since the outgoing interface G3/0 of P2 has reserved a part of bandwidth for the service 2 on PE1, the free bandwidth is not enough to reserve bandwidth for the service 2 on PE3, so the reservation state in this record is specifically a reservation failure, and in the case of the bandwidth reservation failure, the corresponding record in the reserved bandwidth data table may be deleted, i.e. the content recorded in the third row in table 1 may be deleted.

Similarly, the head node device may also maintain its own reserved bandwidth data table in the local memory, and the reserved bandwidth data tables of the head node devices PE1 and PE3 may be shown in table 2 and table 3, respectively:

TABLE 2

；

TABLE 3 Table 3

For the previous example, if PE1 shown in fig. 3 determines SR Policy PE1_1 as the target Policy of service 2, PE1 reserves bandwidth for service 2 on the egress interface G1/0 on the target Policy path, so the contents shown in the second row of table 2 are recorded in the reserved status data table, and the specific meaning may refer to the description of table 1 above.

If PE3 determines SR Policy PE3_1 as the target Policy of service 2, PE3 reserves bandwidth for service 2 on the egress interface G1/0 on the target Policy path, so the content shown in the second row of table 3 is recorded in the reserved status data table, and the specific meaning may refer to the description of table 1.

In the embodiment of the invention, the deterministic probe message carries the deterministic check field, and the deterministic check field comprises a service ID field, a source router ID field, a sink router ID field, a service bandwidth field, a retention time field and/or a reserved state field, so that the intermediate node receiving the deterministic probe message can carry out bandwidth reservation based on the content filled in the deterministic check field, and the process of reserving the first bandwidth for the first service has higher accuracy and certainty.

In an embodiment of the present invention, the step of determining whether there is a target Policy that is reserved successfully by the intermediate node specifically includes:

and the tail node receiving the target Policy feeds back a reverse deterministic message after receiving the deterministic detection message, and judges whether the target Policy which is reserved successfully by the intermediate node exists or not based on a reserved state field in the reverse deterministic message.

Specifically, after receiving the deterministic probe message forwarded by the last hop node, the tail node of the target Policy adds a reverse identifier to the deterministic probe message to obtain a reverse deterministic message, and returns the reverse deterministic message to the head node of the target Policy through a reverse path of a transmission path of the deterministic probe message.

Under the condition that the head node equipment carries out deterministic verification on a plurality of target polices at the same time, the head node equipment can receive a plurality of reverse deterministic messages returned by the target polices, judge whether intermediate nodes in a path of the target Policy corresponding to each reverse deterministic message are reserved successfully or not based on a reserved state field in each reverse deterministic message, and judge whether the target Policy which is reserved successfully by the intermediate nodes exists for the first service or not.

Under the condition that the head node equipment sequentially performs deterministic verification on different target polices, the head node equipment can receive a reverse deterministic message returned by the current target Policy, determine whether all intermediate nodes on a current target Policy path are reserved successfully based on a reserved state field in the reverse deterministic message, if so, directly switch a first service to the current target Policy, and if not, perform deterministic verification on the next hop target Policy until the intermediate nodes are determined to reserve successful target polices, or until SRv Policy in a Policy set corresponding to the first service is not determined to reserve successful target polices after the completion of the uniform verification of the intermediate nodes, and determine that the first service has failed in routing.

As an example, if PE1 in fig. 3 determines SR Policy PE1_1 as the target Policy, the deterministic probe packet sent by PE1 is transmitted through path PE1- > P1- > PE2, after PE2 receives the deterministic probe packet, a reverse identifier is added to the packet and the reverse deterministic packet is sent, where the reverse deterministic packet is transmitted through path PE2- > P1- > PE 1.

After receiving the reverse deterministic message, the intermediate node on the path of the target Policy can be identified as the reverse deterministic message by the reverse identifier in the reverse deterministic message, so that the reverse deterministic message does not play a role in indicating the intermediate node to reserve bandwidth.

Specifically, before the deterministic probe message reaches the tail node of the target Policy, the intermediate nodes in the path of the target Policy update the reserved status field according to whether the node can reserve the first bandwidth for the first service, so after receiving the reverse deterministic message returned by the tail node, the head node device can determine whether the intermediate nodes of the target Policy reserve the first bandwidth for the first service according to the content filled in the reserved status field of the reverse deterministic message.

As an example, to enable the tail node to return the reverse deterministic message to the head node device of the target Policy along the original path, the target Policy may be required to enable BFD (Bidirectional Forwarding Detection ) probing.

Specifically, the existing SRv6 Policy does not maintain its own state through the messages sent by the nodes, so SRv6 Policy mainly completes path fault detection through BFD detection, and implementation of BFD detection requires that the sending path and the return path of the message are consistent.

For a forwarding path, SRv Policy enabling BFD detection is generally configured with two SRv6 policies of bidirectional common paths at a head node and a tail node respectively, and the nodes passed by paths of the two SRv6 policies are consistent, so that BFD detection on the path can be achieved.

Therefore, the embodiment of the invention can utilize the reverse SRv6 Policy configured by the SRv Policy capable of BFD detection at the tail node to realize the return stroke of the reverse deterministic message.

In the embodiment of the invention, the intermediate node updates the reserved state field in the deterministic probe message, the head node equipment judges whether the target Policy which is reserved successfully by the intermediate node exists or not through the reserved state field in the reverse deterministic message returned by the tail node of the target Policy, and multiplexes the sent deterministic probe message, so that the method has higher efficiency and can save calculation resources when judging whether the target Policy which is reserved successfully by the intermediate node exists for the first service or not.

In one embodiment of the present invention, the deterministic probe message is an IFIT message carrying a deterministic check field.

Specifically, in order to ensure that the traffic of the traffic in the SRv network can be forwarded smoothly and scheduled in time, the head node device generally continuously performs IFIT detection on all SRv policies configured on the head node, and obtains path quality of these SRv6 policies. That is, the head-node device will periodically send IFIT messages over these SRv Policy.

That is, the first node device in the SRv network needs to periodically send the IFIT message through SRv6 Policy.

Therefore, the embodiment of the invention can obtain the deterministic probe message by expanding the deterministic check field in the IFIT message. The IFIT message sent by the head node equipment can be used for carrying out deterministic verification on the target Policy on the basis of being used for detecting the path quality of SRv Policy.

Specifically, after the head node device determines one or more SRv policies as target policies of the first service and reserves a first bandwidth for the first service, a deterministic check field may be added in the IFIT message in the next period, and the IFIT message after the deterministic check field is added, that is, a deterministic detection message, is sent through the corresponding target policies, so as to implement deterministic check on the target policies.

As an example, the deterministic check field may be extended in an IFIT message organized in a TLV (Tag-Length-Value) format. For the specific content of the deterministic check field, reference may be made to the description hereinbefore.

In this case, the aforementioned probing period, specifically, the period in which the head node device transmits the IFIT message.

In the embodiment of the invention, the deterministic check field is expanded in the IFIT message to obtain the deterministic detection message which can be used for carrying out deterministic check on the target Policy, and the IFIT message sent by the head node equipment in the process of detecting the path quality of SRv Policy is multiplexed, so that the process of generating and sending the deterministic detection message can be simplified, the cost of processing the message by each node in the path is saved, and the efficiency of carrying out distributed scheduling on the flow is improved.

In one embodiment of the invention, the first bandwidth filled in the traffic bandwidth field is determined based on traffic statistics of the first traffic.

Specifically, a traffic statistic value of the first service under the current SRv Policy may be obtained, so as to obtain the first bandwidth.

As an example, the policy+service ID flow statistics function of the head node device may be started, and the flow teclassfband (service flow bandwidth) of the first service under the current SRv Policy may be obtained in real time.

It should be noted that if the time for switching the first service to the current SRv Policy is less than the acquisition period, the acquired traffic is marked as an outlier, for example, the first bandwidth may be marked as-1, and when the acquired traffic is an outlier, the bandwidth is not reserved for the first service in the path of the target Policy based on the traffic statistics.

Specifically, when the traffic of the first service is just switched to the current SRv Policy, only a part of the traffic of the first service may be counted, and the obtained first bandwidth is smaller than the actual value at this time, so that the traffic statistic value obtained in this case may be determined as an abnormal value, so as to reject an inaccurate traffic statistic value. The acquisition period can be configured according to actual conditions.

As an example, the traffic statistics of the first service in the current SRv Policy may be obtained through an IFIT packet sent by the head-node device.

In the embodiment of the invention, the first bandwidth filled in the service bandwidth field of the deterministic probe message is specifically the traffic statistic value of the first service, so that the first bandwidth is more in line with the bandwidth actually occupied after the traffic of the first service is switched to the target Policy, thereby having higher accuracy when the deterministic check is carried out on the target Policy based on the first bandwidth and better effect of avoiding the scheduling oscillation.

In one embodiment of the present invention, when it is detected that the path quality of the current SRv Policy of one or more second services does not meet the path quality requirement of the second service, and any target Policy of the second service selected from SRv Policy corresponding to the second service is the same as one item of the first service, the distributed traffic scheduling method provided in the embodiment of the present invention further includes:

reserving a second bandwidth for a second service;

for a target Policy that is the same as the first service and the second service, the step of sending a deterministic probe message through the target Policy, so that when determining that the idle bandwidth is not less than the first bandwidth, an intermediate node on a path of the target Policy reserves the first bandwidth for the first service specifically includes:

transmitting a deterministic detection message through a target Policy with the same first service and second service, so that when an intermediate node on a path of the target Policy determines that the idle bandwidth is not less than the first bandwidth and/or the second bandwidth, the intermediate node reserves the first bandwidth for the first service and/or reserves the second bandwidth for the second service;

after reserving the bandwidth for the first service and/or the second service, the method further comprises:

judging whether target polices which are reserved successfully by intermediate nodes exist in target polices corresponding to the second service aiming at the second service; if yes, the second service is switched to the target Policy.

In practical application, the head node device may detect that the path quality of the current SRv Policy of multiple services does not meet the path quality requirement of the corresponding service, and the same target Policy exists in the target policies determined for the multiple services, where the head node device needs to perform deterministic verification on one entry of the target Policy for the multiple services.

In the embodiment of the present application, the head node device may perform deterministic verification on a plurality of services on a target Policy by using the same deterministic probe packet, and the embodiment of the present application is described below by taking a first service and one or more second services as examples.

The process of determining the Policy set corresponding to the second service and performing deterministic verification on the target Policy of the second service by detecting, at the head node device, that the current SRv Policy of the second service does not meet the path quality requirement of the second service may refer to the description of the first service in the foregoing embodiment of the present application, which is not repeated herein.

Specifically, when the head node device detects that the path quality of the current SRv Policy of one or more second services does not meet the path quality requirement of the corresponding service, any target Policy of the second service selected from SRv Policy corresponding to the second service is identical to an entry target Policy of the first service, bandwidths are reserved for the first service and the second service in turn, in the process of performing deterministic verification on the target Policy identical to the first service and the second service, the same deterministic detection message can be sent through the target Policy, and the deterministic detection message can indicate that an intermediate node on the path of the target Policy reserves bandwidth for the first service and/or the second service in turn when the idle bandwidth is not less than the first bandwidth and/or the second bandwidth, and the deterministic verification on the first service and the second service is simultaneously implemented on the target Policy.

The deterministic probe packet may specifically carry a plurality of groups of deterministic check fields, where each group of deterministic check fields corresponds to a service to be checked. Specifically, if deterministic verification is required to be performed on the same target Policy for n services, the deterministic probe message includes n groups of deterministic verification fields, where the maximum value of n, that is, the maximum number of services that can be simultaneously deterministic verified on one target Policy, may be configured according to the performance and network condition of the node device. For the specific content of the deterministic check field, reference may be made to the description hereinbefore.

As an example, if deterministic verification needs to be performed on the same target Policy for n services, n groups of deterministic verification fields may be specifically added in the IFIT packet of the next period.

Referring to fig. 3, if the head node device determines that SR Policy PE1_2 is the target Policy of the first service and the target Policy of the second service at the same time, and PE1 needs to perform deterministic check on SR Policy PE1_2 for both service 1 and service 2, a deterministic check field corresponding to service 1 and a deterministic check field corresponding to service 2 may be added in the IFIT packet in the next period, and the deterministic check field is sent through SR Policy PE 1_2.

In the embodiment of the invention, when the head node equipment needs to carry out deterministic check on the same target Policy for the first service and one or more second services, the intermediate node can be indicated to reserve the first bandwidth for the first service and/or reserve the second bandwidth for the second service when the idle bandwidth is not smaller than the first bandwidth and/or the second bandwidth by sending a deterministic probe message on the target Policy, so that the simultaneous deterministic check on a plurality of services on the same target Policy is realized, and the deterministic check on the services is not needed to be carried out by sequentially sending a plurality of deterministic probe messages through the target Policy. The method and the device can reduce the overhead of processing the message by the intermediate node, improve the efficiency of deterministic verification on the target Policy, and improve the efficiency of distributed scheduling of the flow.

In one embodiment of the present invention, the step of reserving a Policy for each successful handover of the first service to the intermediate node specifically includes:

and determining a target Policy with the highest priority from target policies which are reserved successfully from intermediate nodes based on the priority order of the Color values of the first service and the Color values of the target policies, and switching the first service to the target Policy.

Specifically, each service may configure a set of Color value priority orders, and based on the set of Color value priority orders and the Color value of each target Policy, the priority orders of the plurality of target policies corresponding to the first service may be determined.

Under the condition that the head node equipment performs deterministic verification on a plurality of target polices at the same time, it may be determined that a plurality of intermediate nodes reserve successful target polices, and when the first service is switched to one of the target polices, the first service is switched to the one of the target polices with the highest priority.

Taking fig. 3 as an example, when the head node PE1 detects that SR policy_pe1_3 does not meet the path quality requirement of service 2 and determines SR policy_pe1_1 and SR policy_pe1_2 as the target Policy of service 2, it may simultaneously perform deterministic verification on SR policy_pe1_1 and SR policy_pe1_2, and if the intermediate nodes on SR policy_pe1_1 and SR policy_pe1_2 are reserved successfully, since the priority order of Color values corresponding to service 2 is specifically Color 3>Color 2>Color 1 and the Color values of policy_pe1_1 and SR policy_pe1_2 are respectively 2 and 1,SR Policy_PE1_2 are the highest, service 2 may be switched to SR policy_pe1_2.

Under the condition that the head node equipment sequentially performs deterministic verification on a plurality of target polices, the target polices can be sequentially subjected to deterministic verification according to the priority order of the target polices.

Therefore, when the head node device detects that the current SRv Policy path quality of the first service does not meet the path quality requirement of the first service, firstly, SRv6 Policy meeting the path quality requirement and having the highest priority is selected from SRv Policy corresponding to the first service as a target Policy, and deterministic verification is performed on the target Policy. If the certainty check of the target Policy fails, SRv6 Policy meeting the path quality requirement and having the highest priority can be selected from SRv6 policies of other undetermined verifications in the SRv Policy set of the first node device as a new target Policy, and the new target Policy is subjected to certainty check.

Taking fig. 3 as an example, when the head node device PE1 detects that SR policy_pe1_3 does not meet the path quality requirement of service 2, and determines SR policy_pe1_1 and SR policy_pe1_2 as the target Policy of service 2, it can be known from the foregoing description that the SR policy_pe1_2 has a higher priority than SR policy_pe1_1, so that PE1 may perform deterministic check on SR policy_pe1_2 first, and perform deterministic check on SR policy_pe1_1 again when the deterministic check on SR policy_pe1_2 fails.

In the embodiment of the invention, when the first service is switched, particularly according to the priority order of the Color values of the first service and the Color values of the target Policy, the first service is switched to the one-item-target Policy with the highest priority which is reserved successfully by the intermediate node, so that the traffic of the first service can be switched to the SRv Policy with the higher priority preferentially by applying the distributed traffic scheduling method provided by the embodiment of the invention under the condition that the intermediate node on the path with a plurality of target policies is reserved successfully by the intermediate node, and therefore, the traffic demands of different services can be met to the greatest extent on the basis of avoiding scheduling oscillation.

Fig. 5 is another flow chart of a distributed traffic scheduling method according to an embodiment of the present invention, where the distributed traffic scheduling method is specifically applied to an intermediate node device in a SRv network, and referring to fig. 5, the method specifically may include the following steps:

step S501: receiving a certainty detection message sent by a last hop node in a target Policy path of a first service; when the target Policy is that the current SRv Policy path quality of the first service does not meet the path quality requirement of the first service, the head node equipment selects SRv6 Policy meeting the path quality requirement from SRv6 Policy corresponding to the first service; the deterministic probe message comprises a reserved state field, and the reserved state field is used for indicating whether all the path nodes of the deterministic probe message in the path of the target Policy reserve a first bandwidth for the first service.

Step S502: and judging whether the upstream nodes in the path of the target Policy reserve the first bandwidth for the first service based on the reserved state field in the deterministic probe message, if not, executing step S503, and if so, executing step S504.

Step S503: forwarding the deterministic probe message to a next hop node in the path of the target Policy.

Step S504: under the condition that the idle bandwidth is not smaller than the first bandwidth, reserving the first bandwidth for the first service on an output interface in a path of the target Policy, and forwarding a deterministic probe message to a next hop node in the path of the target Policy; and updating a reserved state field in the deterministic probe message under the condition that the idle bandwidth is smaller than the first bandwidth, and forwarding the updated deterministic probe message to a next hop node in the path of the target Policy.

The specific content of the target Policy and the deterministic probe packet may refer to the description in the embodiment of the distributed traffic scheduling method applied to the head node device.

Specifically, the deterministic check field transmitted through the target Policy includes a reserved status field, where the reserved status field is specifically used to indicate whether all the path nodes of the deterministic probe packet in the path of the target Policy reserve the first bandwidth for the first service, and the intermediate node may determine, based on the field, whether the first bandwidth needs to be reserved for the first service, and whether the field needs to be updated.

After the intermediate node receives the deterministic probe message sent by the last hop node, if any intermediate node in the target Policy path is characterized by the reserved status field, and no matter whether the subsequent node reserves the bandwidth for the first service or not, the deterministic check of the target Policy fails, so that the deterministic probe message can be directly forwarded to the next hop node in the path of the target Policy without reserving the bandwidth for the first service.

If the reserved status fields represent that the intermediate nodes in the target Policy path are reserved bandwidths for the first service, reserving the first bandwidth for the first service on an outgoing interface in the path of the target Policy under the condition that the idle bandwidth is not less than the first bandwidth, forwarding the deterministic probe message to a next hop node in the path of the target Policy, updating the reserved status fields in the deterministic probe message under the condition that the idle bandwidth is less than the first bandwidth, and forwarding the updated deterministic probe message to the next hop node in the path of the target Policy.

As an example, if a path of a certain target Policy of the first service is specifically PE4- > P5- > PE6, and all the path nodes in the path of the target Policy are reserved with the first bandwidth for the first service by filling 1 in the reserved status field, and any path node in the path of the target Policy is not reserved with the first bandwidth for the first service by filling 0. Under the condition that the idle bandwidth of the PE4 is not smaller than the first bandwidth, the PE4 sends a deterministic probe message through the target Policy, at the moment, a reserved state field of the deterministic probe message is filled with 1, after receiving the deterministic probe message, the next hop node P4 determines that the bandwidth needs to be reserved for the first service, when the idle bandwidth is not smaller than the first bandwidth, the bandwidth is reserved for the first service, and the deterministic probe message is forwarded to the next hop node P5; when the idle bandwidth is smaller than the first bandwidth, 1 filled in a reserved state field of the deterministic probe message is required to be updated to 0, the updated deterministic probe message is sent to P5, and when the deterministic probe message is received, the P5 determines that the bandwidth is not required to be reserved for the first service any more, and the message is directly sent to a next hop node.

In the embodiment of the invention, the intermediate node in the SRv network specifically judges whether the first bandwidth needs to be reserved for the first service or not based on the reserved status field in the deterministic message sent by the last hop node, and judges whether the reserved status field needs to be updated according to the reserved condition of the node under the condition that the first bandwidth needs to be reserved for the first service, so that the deterministic checking process performed on the target Policy has higher efficiency and accuracy.

In one embodiment of the present invention, the distributed traffic scheduling method further includes:

the method comprises the steps that a tail node in a path of a target Policy receives a reverse deterministic message fed back after receiving a deterministic detection message, and whether intermediate nodes in the path of the target Policy reserve the first bandwidth for a first service or not is judged based on a reserved state field in the reverse deterministic message;

Specifically, there may be one or more intermediate nodes in the path of the target Policy, and the idle bandwidth on each node in the path of the target Policy may be different, so in the process that each node reserves the first bandwidth for the first service, a situation may occur that the first-half node reserves the first bandwidth for the first service in the path, but the second-half node does not reserve the first bandwidth for the first service. Therefore, in order to avoid that the first bandwidth reserved by the first half node on the path of the target Policy affects normal transmission of other traffic flows, after the intermediate node device receives the reverse deterministic message, if the reserved status field in the reverse deterministic message indicates that any intermediate node does not reserve the first bandwidth, the intermediate node device that has reserved the first bandwidth for the first service may release the portion of the first bandwidth, and delete the corresponding record in the local reserved bandwidth data table.

As an example, if a path of a certain target Policy of the first service is specifically PE4- > P5- > PE6, if P4 reserves a first bandwidth for the first service and the free bandwidth on P5 is smaller than the first bandwidth, P5 fails to reserve the bandwidth for the first service, and at this time, the reserved status field in the reverse deterministic packet returned by PE6 specifically indicates that any intermediate node in the path of the Policy does not reserve the bandwidth for the first service, so that after receiving the reverse deterministic packet, P4 may release the first bandwidth reserved for the first service.

In the embodiment of the invention, after receiving the reverse deterministic message returned by the tail node, the intermediate node equipment on the path of the target Policy releases the part of bandwidth if the reserved state field in the reverse deterministic message indicates that any intermediate node does not reserve the first bandwidth and the node reserves the first bandwidth, so that the bandwidth utilization rate in the path can be ensured on the basis of carrying out deterministic verification on the target Policy, and the influence on the normal transmission of other traffic is avoided.

Fig. 6 is a schematic flow chart of a distributed traffic scheduling method according to an embodiment of the present invention, where the distributed traffic scheduling method is specifically applied to an inter-tail node device in a SRv network, and referring to fig. 6, the method may specifically include the following steps:

Step S601: after receiving a determination detection message sent by a head node device in a target Policy path through the target Policy, adding a reverse identifier in the determination detection message to obtain a reverse determination message; and when the target Policy is that the current SRv6 Policy path quality of the first service does not meet the path quality requirement of the first service, the head node equipment selects SRv6 Policy meeting the path quality requirement from SRv6 Policy corresponding to the first service.

Step S602: and returning a reverse deterministic message through the path of the target Policy, wherein the reverse deterministic message is used for informing an upstream node in the path of the target Policy whether intermediate nodes in the path of the target Policy reserve a first bandwidth for the first service.

Specifically, the deterministic probe message includes a reserved status field, and the reverse identifier added by the tail node device is used to enable the upstream node in the target Policy path to identify the message as a reverse deterministic message, so as to determine whether intermediate nodes in the path of the target Policy reserve the first bandwidth for the first service based on the reserved status field therein.

In the embodiment of the invention, the intermediate node updates the reserved status field in the deterministic probe message, and the tail node equipment adds the reverse identifier in the deterministic probe message and returns the obtained reverse deterministic message, so that the aim of informing the upstream node in the target Policy whether the intermediate nodes in the path of the target Policy reserve the first bandwidth for the first service is realized, and the deterministic probe message is multiplexed, so that the head node equipment has higher efficiency and can save calculation resources when judging whether the intermediate nodes reserve the successful target Policy for the first service.

Based on the same inventive concept, according to the distributed traffic scheduling method provided by the above embodiment of the present invention, the embodiment of the present invention further provides a distributed traffic scheduling device, where the device is applied to a head node device in a SRv network. As shown in fig. 7, the apparatus includes:

a selecting module 701, configured to select, when it is detected that the current SRv Policy path quality of the first service does not meet the path quality requirement of the first service, a Policy set that meets the path quality requirement from SRv policies corresponding to the first service; the Policy set contains at least one target Policy;

A reservation module 702, configured to reserve a first bandwidth for a first service for each target Policy, and send a deterministic probe packet through the target Policy, so that an intermediate node on a path of the target Policy reserves the first bandwidth for the first service when determining that an idle bandwidth is not less than the first bandwidth;

a first judging module 703, configured to judge whether there is a target Policy that is reserved successfully by the intermediate node;

and a switching module 704, configured to switch the first service to an entry Policy reserved successfully by the intermediate node if the determination result of the first determining module 703 is that the first service exists.

In the distributed traffic scheduling device provided by the embodiment of the invention, when the head node equipment detects that the current SRv Policy path quality of the first service does not meet the path quality requirement of the first service, a Policy set meeting the path quality requirement is selected from SRv policies corresponding to the first service, before the first service is switched, a first bandwidth is reserved for the first service for each target Policy, and a determination detection message is sent through the target Policy, so that when the idle bandwidth is not less than the first bandwidth, the first bandwidth is reserved for the first service by an intermediate node on the path of the target Policy, then whether the target Policy which is reserved by the intermediate node is existed or not is judged, and when the first service is switched, the first service is switched to the first item of Policy which is reserved by the intermediate node.

the service ID field is used for indicating the type of the service;

a source router ID field for indicating a head node of the target Policy;

a sink router ID field for indicating a tail node of the target Policy;

In one embodiment of the present invention, the first determining module 703 is specifically configured to:

In one embodiment of the present invention, the reservation module 702 is specifically configured to:

In one embodiment of the present invention, the selection module 701 includes a selection unit, specifically configured to:

In one embodiment of the present invention, the selecting module 701 detects that the path quality of the current SRv Policy of one or more second services does not meet the path quality requirement of the second service, and any target Policy of the second service selected from SRv Policy corresponding to the second service is the same as one item of the first service;

The selection module 701 is further configured to:

reserving a second bandwidth for a second service;

for the target Policy that the first service and the second service are the same, the reservation module 702 is specifically configured to:

the switching module 704 is further configured to:

Based on the same inventive concept, according to the distributed traffic scheduling method provided by the above embodiment of the present invention, the embodiment of the present invention further provides a distributed traffic scheduling device, where the device is applied to an intermediate node device in a SRv network. As shown in fig. 8, the apparatus includes:

a receiving module 801, configured to receive a deterministic probe packet sent by a last hop node in a path of a target Policy of a first service; when the target Policy is that the current SRv Policy path quality of the first service does not meet the path quality requirement of the first service, the head node equipment selects SRv6 Policy meeting the path quality requirement from SRv6 Policy corresponding to the first service; the deterministic probe message comprises a reserved state field, wherein the reserved state field is used for indicating whether all the path nodes of the deterministic probe message in a path of the target Policy reserve a first bandwidth for a first service;

A second judging module 802, configured to judge, based on the reserved status field in the deterministic probe packet, whether the upstream nodes in the path of the target Policy reserve the first bandwidth for the first service;

a forwarding module 803, configured to forward the deterministic probe packet to a next hop node in the path of the target Policy if the determination result of the second determining module is negative; if the judgment result of the second judgment module is yes, reserving a first bandwidth for the first service on an output interface in a path of the target Policy under the condition that the idle bandwidth is not smaller than the first bandwidth, and forwarding a deterministic probe message to a next hop node in the path of the target Policy; and updating a reserved state field in the deterministic probe message under the condition that the idle bandwidth is smaller than the first bandwidth, and forwarding the updated deterministic probe message to a next hop node in the path of the target Policy.

In one embodiment of the invention, the apparatus further comprises:

the third judging module is used for receiving a reverse deterministic message fed back by a tail node in the path of the target Policy after receiving the deterministic detection message, and judging whether intermediate nodes in the path of the target Policy reserve a first bandwidth for the first service or not based on a reserved state field in the reverse deterministic message;

the releasing module is used for releasing the first bandwidth under the condition that the first bandwidth is reserved for the first service if the judging result of the third judging module is negative; and if the judgment result of the third judgment module is yes, releasing the first bandwidth reserved for the first service after the target Policy is switched for the first service.

Based on the same inventive concept, according to the distributed traffic scheduling method provided by the above embodiment of the present invention, the embodiment of the present invention further provides a distributed traffic scheduling device, where the device is applied to a tail node device in a SRv network. As shown in fig. 9, the apparatus includes:

the adding module 901 is configured to add a reverse identifier to the deterministic probe packet after receiving a deterministic probe packet sent by a first node device in a target Policy path through the target Policy, so as to obtain a reverse deterministic packet; and when the target Policy is that the current SRv6 Policy path quality of the first service does not meet the path quality requirement of the first service, the head node equipment selects SRv6 Policy meeting the path quality requirement from SRv6 Policy corresponding to the first service.

And a return module 902, configured to return, through the path of the target Policy, a reverse deterministic packet, where the reverse deterministic packet is configured to notify, to an upstream node in the path of the target Policy, whether intermediate nodes in the path of the target Policy all reserve the first bandwidth for the first service.

The embodiment of the invention also provides an electronic device, as shown in fig. 10, which comprises a processor 101, a communication interface 102, a memory 103 and a communication bus 104, wherein the processor 101, the communication interface 102 and the memory 103 complete communication with each other through the communication bus 104,

A memory 103 for storing a computer program;

the processor 101 is configured to execute a program stored in the memory 103, and implement the following steps:

when detecting that the current SRv6 Policy path quality of the first service does not meet the path quality requirement of the first service, selecting a Policy set meeting the path quality requirement from SRv6 policies corresponding to the first service; the Policy set contains at least one target Policy;

reserving a first bandwidth for a first service according to each target Policy, and sending a deterministic probe message through the target Policy, so that when an intermediate node on a path of the target Policy determines that the idle bandwidth is not less than the first bandwidth, reserving the first bandwidth for the first service; judging whether target Policy which is reserved successfully by the intermediate nodes exists or not;

The communication bus mentioned above for the electronic devices may be a peripheral component interconnect standard (Peripheral Component Interconnect, PCI) bus or an extended industry standard architecture (Extended Industry Standard Architecture, EISA) bus, etc. The communication bus may be classified as an address bus, a data bus, a control bus, or the like. For ease of illustration, the figures are shown with only one bold line, but not with only one bus or one type of bus.

The communication interface is used for communication between the electronic device and other devices.

The Memory may include random access Memory (Random Access Memory, RAM) or may include Non-Volatile Memory (NVM), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the aforementioned processor.

The processor may be a general-purpose processor, including a central processing unit (Central Processing Unit, CPU), a network processor (Network Processor, NP), etc.; but also digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), field programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

In yet another embodiment of the present invention, a computer readable storage medium is provided, in which a computer program is stored, which when executed by a processor, implements the steps of any of the above-described distributed traffic scheduling methods.

In yet another embodiment of the present invention, a computer program product containing instructions that, when run on a computer, cause the computer to perform any of the distributed traffic scheduling methods of the above embodiments is also provided.

In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, produces a flow or function in accordance with embodiments of the present invention, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in or transmitted from one computer-readable storage medium to another, for example, by wired (e.g., coaxial cable, optical fiber, digital Subscriber Line (DSL)), or wireless (e.g., infrared, wireless, microwave, etc.). The computer readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid State Disk (SSD)), etc.

It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.

In this specification, each embodiment is described in a related manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for distributed flow scheduling apparatus, electronic devices, and computer readable storage medium embodiments, since they are substantially similar to method embodiments, the description is relatively simple, and references to the parts of the description of method embodiments are only needed.

The foregoing description is only of the preferred embodiments of the present invention and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.

Claims

1. A distributed traffic scheduling method, applied to a head node device in an IPv6 based segment routing SRv6 network, comprising:

2. The method according to claim 1, wherein the deterministic probe message carries a deterministic check field, the deterministic check field comprising a traffic ID field, a source router ID field, a sink router ID field, a traffic bandwidth field, a retention time field and/or a reservation status field;

the service ID field is used for indicating the type of the service;

3. The method according to claim 2, wherein the step of determining whether there is a target Policy that is successfully reserved by the intermediate nodes includes:

4. The method of claim 2, wherein the first bandwidth filled in the traffic bandwidth field is determined based on traffic statistics of the first traffic.

5. The method of claim 2, wherein the deterministic probe message is a flow-along detection IFIT message carrying the deterministic check field.

6. The method according to claim 1, wherein the step of reserving a first bandwidth for the first traffic for each of the target polices and transmitting a deterministic probe message through the target Policy to cause an intermediate node on a path of the target Policy to reserve the first bandwidth for the first traffic when determining that an idle bandwidth is not less than the first bandwidth comprises:

7. The method of claim 1, wherein the step of reserving a successful Policy for each of the first traffic to the intermediate node comprises:

8. The method as recited in claim 1, further comprising:

the method further comprises the steps of:

9. A distributed traffic scheduling method, applied to an intermediate node device in a SRv network, comprising:

10. The method as recited in claim 9, further comprising:

11. A distributed traffic scheduling method, applied to a tail node device in a SRv network, comprising:

after receiving a deterministic probe message sent by a head node device in a target Policy path through the target Policy, adding a reverse identifier in the deterministic probe message to obtain a reverse deterministic message; when the target Policy is that the current SRv6 Policy path quality of the first service does not meet the path quality requirement of the first service, the head node device selects SRv6 Policy meeting the path quality requirement from SRv6 Policy corresponding to the first service;

12. A distributed traffic scheduling apparatus, for use in a head-node device in a SRv network, comprising:

13. A distributed traffic scheduling apparatus, for use in an intermediate node device in a SRv network, comprising:

14. A distributed traffic scheduling apparatus, for use in a tail node device in a SRv network, comprising:

the adding module is used for adding a reverse identifier into the deterministic detection message after receiving the deterministic detection message sent by the head node equipment in the target Policy path through the target Policy to obtain a reverse deterministic message; when the target Policy is that the current SRv6 Policy path quality of the first service does not meet the path quality requirement of the first service, the head node device selects SRv6 Policy meeting the path quality requirement from SRv6 Policy corresponding to the first service;

15. The electronic equipment is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory are communicated with each other through the communication bus;

a memory for storing a computer program;

a processor for carrying out the method steps of any one of claims 1 to 8, or 9 to 10 or 11 when executing a program stored on a memory.

16. A computer-readable storage medium, characterized in that the computer-readable storage medium has stored therein a computer program which, when executed by a processor, implements the method steps of any of claims 1-8, or 9-10 or 11.